Edge Sound raises $2M to give physical AI an audio layer

The company’s starting point was sports. When a basketball hits the court, fans are stomping in the stands and bass pounds through arena speakers; microphones capture all of it at once. For an AI system trying to isolate the dribble, those overlapping sounds make the task difficult. Edge Sound’s approach connects audio with tracking and visual data to determine which sound belongs to which object.

“The issue in audio is it’s not object oriented; it’s channel oriented,” co-founder and Chief Executive Valtteri Salomaki told SiliconANGLE in an interview.

For broadcasters, that connection means following a player’s sounds through a crowded arena. Edge is betting the same approach can help robots and other physical AI systems interpret their surroundings, including activity their cameras cannot yet see. Humans do this automatically because sight, sound and physical sensation arrive at once, but in physical AI, sound has been treated as secondary to vision because cameras are already widespread. Salomaki said audio arrives in channels: microphones positioned around a space or on objects listen at locations and produce signals that identify sounds, but those signals do not directly associate with anything until another system turns them into identifiable information.

“It’s split into two layers,” Salomaki said. “One is how do you create a digital twin of all sensors. The other one is how do you create AI machine learning models to source separate sound elements.”

In the sports arena demonstration, the system uses microphone locations, pickup patterns and player tracking to select useful audio, then associates locations; visual and audio are paired to identify objects such as players, the ball, a coach or a referee with a whistle. Users can import 3D spaces or 2D layouts with coordinates and place sensors within that representation, creating a spatial map of the environment and its sensors—the “digital twin” Salomaki described.

Other possible uses include autonomous vehicles, which could be given digital ears to hear whether a siren is approaching. Humans identify sound and direction because our ears sit on the sides of the head; a sound from the left is louder on that side even when the source is not visible. The Doppler effect, in which pitch rises or falls with speed, can give an autonomous vehicle an immediate warning that an emergency vehicle is approaching from a blind corner, and its direction, even if the lights are not visible to cameras.

Edge develops a general model and fine-tunes it for customer needs. These models face different delay targets for real-time activity; Salomaki cited a sub-33-millisecond tolerance for sports broadcasting and said other environments might tolerate 100 or 200 milliseconds. Useful perception needs to arrive on time for most applications because hearing what happens needs to be paired with seeing.

The company is building toward hardware independence with its digital twin software and AI models, allowing it to assemble new applications around a core technology and become audio infrastructure that works alongside the microphones and cameras customers already use for physical AI robots, media and other systems. With the funding, Salomaki said, Edge will complete integration with existing enterprise customers and make the platform’s components transferable across industries ahead of a Series A.

Live sports has already given Edge a commercial market as it developed and validated its perception technology and used it to train and build models. Its Virtual Sound Engine, which automatically captures and isolates sound from complex environments, is already deployed commercially across the industry. Customers and partners include American audio equipment manufacturer Shure Inc., the National Basketball Association, Fox Sports Media Group and the United States Tennis Association Inc.